1
The reflection of historical language contact in present-day Dutch and Swedish
Charlotte Gooskens, Renée van Bezooijen and Sebastian Kürschner
University of Groningen
[email protected], [email protected], [email protected]
Charlotte Gooskens, Renée van Bezooijen and Sebastian Kürschner
University of Groningen
Scandinavian Department
Postbus 716
9700 AS Groningen
The Netherlands
2
Abstract
In the present study we quantitatively examined similarly constructed samples of
formal spoken Swedish and Dutch in order to compare the composition of the
lexicons. Results showed that Swedish has many more loans than Dutch, namely
44.4% against 27.9%. Within the Swedish loans there is a large compartment of Low
German (38.7%), whereas most loans in Dutch have a French origin (63.8%). The
differences in terms of the number and distribution of loanwords between the lexical
profiles of Swedish and Dutch appear to be stable, as they were attested both in the
present study and in previous studies. They can be attributed to differences in the
linguistic distances between source and borrowing languages and to differences in the
intensity of the contacts.
3
The reflection of historical language contact in present-day Dutch and Swedish
Charlotte Gooskens, Renée van Bezooijen and Sebastian Kürschner
University of Groningen
1. Introduction
Globalization can literally be defined as the process by which local or regional things
and phenomena are transformed into global ones. Globalization may affect the world
in a large number of ways, e.g. industrial, technical, economic, political, ecological,
social, and linguistic. At present, linguistic globalization is particularly evident in the
all-pervasive influence of English on the worlds’ languages. However, linguistic
globalization has occurred in previous centuries as well. In fact, according to
Thomason (2001: 8) “there is no evidence that any languages have developed in total
isolation from other languages”. A few pages further she contends “language contact
is the norm, not the exception” (p.10).
It is well known that language contact may result in linguistic changes at all
levels: phonetic, phonological, morphological, syntactical, prosodic and lexical,
depending, among other things, on the intensity of the contact and the degree of
relatedness of the languages involved. In this article we will focus on the lexical level,
which is generally assumed to be the level that is most easily influenced (Thomason
4
2001: 69). We were interested to know to what extent different language contact
histories may lead to differences in the composition of the lexicon of present-day
languages. We opted for comparing Swedish and Dutch. These two Germanic
languages share many stems due to their common origin in Proto-Germanic. So,
originally their lexicons were very similar. However, the two languages have diverged
considerably, as a consequence of both language internal and language external
factors, in particular language contact. The nature of these contacts is well
documented. In the course of time, Dutch and Swedish have been in contact with the
same languages, particularly Low and High German, Latin and French, but the
intensity and duration of these contacts differed considerably (see Section 2). It is not
our intention to throw new light on these historical developments. Our research
question is: How are the similarities and differences in language contact in the past
reflected in the Dutch and Swedish languages as they are used at present?
The few studies that have assessed the composition of the present-day Dutch and
Swedish lexicons in a quantitative way were exclusively based on newspaper texts
(see Section 3). Moreover, different methodologies and different types of texts were
used, which makes it difficult to compare the results. For our contrastive
investigation, we took great care to ensure that the databases for the two languages
were constructed in exactly the same manner. Both databases include prepared
speeches and spontaneous dialogues that can be characterized as formal. All material
originates from meetings held in the European Parliament in the first months of 2000,
either in Dutch or in Swedish. In this way, the content and style of the speech material
was kept constant. In view of the setting (monologues and dialogues for a large
public) and topics (politics, economics, administration), the style can be characterized
5
as formal. Moreover, for both Dutch and Swedish we only looked at the most frequent
words, as the analysis of frequent words will be less prone to chance fluctuation than
the analysis of infrequent words. Finally, the same procedure of word selection and
coding was applied to both languages (see Section 4). All this taken together, we
provide a quantitative study based on a large and reliable data set that is suitable for a
valid comparison of the two languages involved.
Our research questions can be formulated as follows:
(1) What are the proportions of inherited words and loanwords in contemporary
Dutch and Swedish?
(2) What are the origins of the loanwords in the two languages?
(3) Which historical prerequisites, such as language contact situation or linguistic
distance, can help to explain the differences between the lexical profiles?
2. Historical background
The histories of loanwords in Dutch and Swedish reveal similarities as well as
differences.1 In the early Middle Ages, both languages borrowed many Latin and
1 For the history of loans in Dutch, cf. e.g. van der Sijs (2005), for Swedish cf. Edlund & Hene
(2004).
6
some Greek words as part of Christianization. Throughout the Middle Ages, Latin
remained influential because of its leading role in the church and in the sciences.
In the late Middle Ages, both the Swedish and Dutch dialects had intensive
contacts with Low German. This was the language of the Hanseatic League, which
constituted a strong economic power. The Hanseatic merchants, who were located in
Northern Germany (mainly Lübeck and Hamburg), built up a trade network covering
all of Northern Europe. As a consequence of the closer relationship between West
Germanic Dutch and Low German, the dialects which would later constitute the basis
of Dutch shared many words with Low German. The dialect contacts in the Hanseatic
era may have changed the frequency of use of some native words, but this did not lead
to intense borrowing. In contrast, North Germanic Swedish, with fewer parallels in
the lexicon, was altered by the Low German influence, which is evident from a large
number of loanwords.
Already in the early medieval period the Dutch dialects were strongly influenced
by French (Van der Sijs 1996: 134). This can be inferred from the great number of
loans that go back to Old French. Examples are vijg ‘fig’ from O.Fr. figue, kussen
‘pillow’ from O.Fr. cussin or coissin, prei ‘leaks’ from O.Fr. porée, aalmoes ‘alms’
from O.Fr. almosne, toren ‘tower’ from O.Fr. tur and fel ‘fierce’ from O.Fr. fel. At
this point of time, the centre of economic power in the Dutch-speaking dialect area
was located in the Southern Flemish part, which was strongly influenced, both
economically and politically, by the neighboring French-speaking area. The influence
of French on Swedish started much later, in the seventeenth and eighteenth centuries.
In both Dutch and Swedish the dominant position of French remained stable until the
nineteenth century. The upper classes were bilingual, and used many French words in
7
their native language. As their manner of speaking had prestige, the French words
were adopted by the middle and lower classes, and thus got incorporated into the
general language (Van der Sijs 1996: 139).
The standardization of the West and North European written languages started in
the sixteenth century. With the translation of the Luther bible from High German into
Dutch in the first half of the sixteenth century and with a considerable number of
Germans living in the Dutch language area, High German gained influence on Dutch.
From the time of the reformation and the Thirty Years’ War (1618-1648) onwards, a
comparable development took place in Swedish. In the nineteenth century, High
German influence was at its top, especially in the domains of science, industry and
trade. In the twentieth century, and especially after the Second World War, English
words started to be adopted in Dutch and Swedish, especially within the domains of
industrialization, transport, technology and sports. In fact, English is now almost the
sole provider of loanwords.
In contrast to Swedish, the Dutch language has (partly) been shaped by its
colonial history, specifically in Indonesia. This influence was not as strong, however,
as that of the European languages mentioned above.
In Figure 1 we present a schematized overview of the intensity and duration of the
language contacts for Dutch and Swedish. Low and High German are represented by
the same line. Low German contacts mainly took place in the Middle Ages (before
1550), whereas High German contacts mainly occurred from early modern times
(after 1550). The two contact situations can thus easily be distinguished. The contacts
with Low German-speaking Hanseatic merchants from Northern Germany involved
8
the entire population in the neighboring countries. The use of High German, in
contrast, was largely restricted to the court and institutions of higher education.
Moreover, whereas the contact with Low German was mainly established via the
spoken language, the contact with High German took place mainly via the written
channel of communication (cf. Braunmüller 2004: 23). This is why we consider the
former to have been more pervasive, affecting larger portions of the population than
the latter. The Dutch and Swedish contacts with Low and High German were
comparable in intensity.
The contact between Dutch and French was more intensive than that between
Swedish and French. French was only a court language in Sweden, whereas it was a
high prestige neighboring language of the Dutch language area. The duration of
intensive contact between Dutch and French was much longer than between Swedish
and French, since it started already in the Middle Ages. For Dutch, the contact with
French was intensive on the written as well as on the spoken level, and for large parts
of the population. In Swedish, by contrast, the contact was mainly written. The use of
spoken French was restricted to a small minority, specifically the influential, highly
educated parts of the population.
Both Dutch and Swedish had contact with Latin. In both languages, this contact
took place on the written level and was therefore only accessible to a small part of the
society for a long period of time. The intensity of this language contact therefore
seems comparable for Dutch and Swedish, as indicated in Figure 1.
Figure 1 here
9
3. Previous investigations
Van der Sijs (1996: 65) presents a small exploratory study of the first four pages of
NRC Handelsblad (one of the major daily newspapers in the Netherlands) of April 7,
1994, totaling 11,872 words. After (1) removing personal and geographic names, (2)
collapsing conjugated verb forms, and (3) splitting up compounds, she retained 2,144
different lexemes. 69.3% of these were inherited words, dating back to the time when
the Germanic languages still formed a unity, and 30.7% were loanwords. Apparently,
Van der Sijs is quite impressed by the high number of loans, for she states that “Dutch
has received, and is still receiving, loanwords warmly and hospitably” (our
translation). The large majority of the loans in the NRC-sample have a Latin, French,
Italian or Spanish origin (82.0%). High and Low German contributed 6.8% of the
loans and English 7.4%. The few remaining loans, totaling 3.7%, are from Greek,
Arabic, Hebrew, Turkish and Celtic.
Gellerstam (1973) made an analysis of the 6,000 most frequent word forms found
in a Swedish frequency dictionary (Allén 1970). This dictionary is based on one
million words from five different Swedish newspapers from 1970. 42.2% of the words
in the texts were inherited words, whereas 47.2% consisted of loanwords. The rest
were words of unknown origin (1.6%) or words from word classes which Gellerstam
excluded from his analysis, i.e. names, numbers etc. (9.1%). Most loans had a High or
Low German (51.1%) or a Latin/Greek (43.0%) origin. Only few loans originated
from French (2.3%) or English (0.8%) or other languages (2.8%).
A comparison of the data reported for Dutch by Van der Sijs (1996) and for
Swedish by Gellerstam (1973) suggests that the Swedish lexicon contains many more
10
loans than Dutch (47.2% versus 30.7%) and that the extent to which different
languages have contributed to the lexicons of the two languages differs. Swedish
appears to have borrowed more from High and Low German than Dutch has (51.1%
versus 6.8% of the loanwords), whereas Dutch appears to have borrowed more from
Romance languages than Swedish (82.0% versus less than 45.3% of the loanwords).
However, this conclusion can only be tentative, as the two samples of newspaper texts
differ in several respects. The Dutch database is small and a mixed sample of frequent
and infrequent words, whereas the Swedish database is large and restricted to high-
frequency words. Moreover, the Dutch database comprised lexemes (with compounds
broken up into their constituent elements), whereas the Swedish sample is based upon
word forms. Moreover, there is a time lapse of more than twenty years between the
two samples.
4. Method
4.1. Material
For our investigation we made use of the so-called Europarl corpus, which can be
downloaded from the internet.2 This is a parallel corpus that is available for eleven
European Community languages, including Swedish and Dutch. Each language is
represented by approximately 28 million words. The corpus consists of monologues
and dialogues by speakers and chairpersons, collected during meetings in the
European Parliament. Both the original speech and the simultaneous translations by
2 http://people.csail.mit.edu/koehn/publications/europarl/
11
the interpreters into the various languages are included. In this way, the same texts
can be selected. This is important for our purpose, since we wanted to compare the
lexical profiles of Dutch and Swedish keeping subject matter and style constant. For
the present study, we selected all the words from the meetings that were held between
January 17 and March 17, 2000 for either language, which sums up to roughly one
million words per language.
For our investigation we used the 1,500 most frequent words in the one-million-
word databases of each of the two languages. The frequency data were gathered for
lexemes rather than word forms. This means that the frequencies of, for example, huis
‘house’ and huizen ‘houses’ were added together. In the frequency database, the
lexeme is represented by the singular form huis. Verbs are represented by the
infinitive forms and adjectives by the undeclined form. In the Europarl corpus, the
Dutch words had already been lemmatized. The Swedish words we lemmatized
ourselves by means of The Granska tagger from The Royal Institute of Technology in
Sweden.3 Out of the 1,500 most frequent words we removed all personal names,
geographical references and interjections. Compounds constitute a problem, as they
may contain words with different etymologies. This is why we split up all transparent
compounds into their simple stems. In this way each part could be categorized
separately. Examples of such compounds in Dutch are badkamer ‘bath room’ and
vrijdagavond ‘Friday night’. Swedish examples are fredsprocess ‘peace process’ and
arbetsgrupp ‘working group’. In accordance with the procedure adopted by Van der
Sijs (1996, see Section 3), compounds with a preposition as the first element were not
3 The Granska tagger is available for download online via
http://www.csc.kth.se/tcs/humanlang/tools.html (accessed December 11, 2006).
12
split up (e.g. Dutch opbellen ‘call’ and Swedish avsluta ‘finish’). Eventually 1,400
Dutch and 1,418 Swedish lexemes remained for further analysis.
4.2. Coding
Each word was given codes that contained the following information:4
(1) Inherited word or loanword
(2) For loanwords: language from which it has been borrowed directly
For Dutch, the etymological information in (1) and (2) was taken from Van der Veen
& Van der Sijs (1997). If the information was not found there, Van der Sijs (1996)
was consulted. The Swedish information was found in Wessén (1960) and Hellquist
(1980).
5. Results and interpretation
5.1. The proportion of inherited words and loanwords
In Figure 2, the percentages of inherited words and loanwords are shown for Dutch
and for Swedish. In Figure 3, the percentages of loanwords in our investigation are
4 The following information was added as well, but not used in the present investigation: original
language, year of introduction into the language, word class, pronunciation, word length, cognate/non-
cognate.
13
compared with the percentages reported in previous investigations. The percentage of
loanwords in Dutch (27.9%) is similar to the percentage which Van der Sijs (1996)
found in a small newspaper corpus from 1994 (30.7%). Also the percentage of
Swedish loanwords (44.4%) is similar to the percentage of loanwords found in the
previous investigation by Gellerstam (1973) in a newspaper corpus (47.2%). So for
both languages the distribution of inherited words and loanwords is almost identical in
two types of formal speech, namely written language in newspapers and monologues
and dialogues from the European parliament. Apparently, the composition of the
present-day Dutch and Swedish lexicons is very stable in this respect. Both the
previous studies and our own study show that Swedish has more loanwords, and
consequently fewer inherited words, than Dutch. To gain insight into the nature of this
difference, we looked at the quantitative contribution of the source languages in closer
detail.
Figure 2 here
Figure 3 here
5.2. The contribution of different languages and the relation with language contact
In Figure 4, the origin of the loanwords in Dutch and Swedish is broken down for
source language. Two differences stand out. First, Swedish has a large percentage of
Low German loans (38.7%) whereas Dutch has none. Second, Dutch has many more
14
French loanwords (63.8%) than Swedish (14.6%). The other differences are much
smaller. For example, Swedish has slightly more Latin and Greek loans than Dutch
(differences of 2.6% and 4.3%, respectively). The number of High German loans is
higher in Swedish as well (a difference of 5.1%).
Figure 4 here
Let us at first take a look at the German loans. In Section 2, we reported that both
language communities underwent considerable influence of Low German due to the
intense contacts with the Hanseatic tradesmen in the Middle Ages. According to our
results, however, this parallel contact situation did not result in equal amounts of
lexical borrowing. While in the Swedish sample Low German constitutes the largest
group of loans (38.7%), Dutch has no Low German words, or at the most one. In fact,
of the 36 Dutch words with a German origin, 35 are attributed unambiguously to High
German, whereas the precise origin of one word, namely grens ‘border’, is unclear. It
could either go back to the Low German grenize or to the High German grenze.
The difference between the numbers of Low German loanwords in the two
languages can be explained when language distances are taken into account. While
Middle Swedish, as a North Germanic language, already diverged considerably,
structurally as well as lexically, from Middle Low German, Middle Dutch and Middle
Low German were part of the same dialect continuum, where mutual intelligibility
was highly probable (Goossens 2000). This holds for the grammar as well as for the
lexicon. As Dutch and Low German were so similar, there was little room for
borrowing. As we pointed out in Section 2, the intense dialect contact is likely to have
15
changed the frequency of single words that were part of both the Low German and
Dutch vocabularies. These words are often unidentifiable, though. We are also
confronted with the problem that it is difficult to establish the etymology of some
words from the available sources of Dutch language history. When a word is found in
Dutch as well as in Low German and High German documents from the Middle Ages,
it is highly probable that this word is originally a West Germanic word. It cannot be
completely excluded, however, that it was introduced to one of the regions only later
as a loan. Due to these difficulties, the true origin of many loans from Low German
may be concealed in Dutch because we interpret them incorrectly as common West
Germanic: “The agreement between Low German and Dutch sometimes makes it
difficult to decide whether a word is borrowed or related” (Van der Sijs 1996: 231;
our translation).
The contact situation in Sweden was one of structurally related languages as well.
Braunmüller (1995) even assumes that semi-communication, i.e. a situation where
languages are so alike that their speakers can communicate each using their own
language, may have been possible between Scandinavians and Low Germans at this
point of time. Nevertheless, the lexical differences between Middle Swedish and
Middle Low German were much larger than between Middle Dutch and Middle Low
German. This makes it easier to identify Low German loans in modern Swedish.
The high number of Low German loans in our Swedish database can also partly
be explained by the loan of derivations: Through prefixation in Low German, the
same roots can appear in different lexemes, which were then borrowed into Swedish –
like the root sluta in ansluta ‘connect’, avsluta ‘finish’, and besluta ‘decide’. Of the
69 Swedish verb stems of Low German origin in our database, only 55 bear different
16
roots. As Diercks (1995) has shown, affixes of this kind have even become modestly
productive in Swedish.
Swedish has fewer High German (14.3%) than Low German loans (39%), but
still more than Dutch (9.2%). This difference can be attributed, at least partly, to the
same factor that we mentioned above to explain the differences for Low German,
namely the larger linguistic distance between Swedish and Low German than between
Dutch and Low German. Moreover, it should be noted that the High German
influence, which mainly took place in the seventeenth and nineteenth centuries, was
less intense than that of Low German in Swedish.
With respect to French and Latin loans, the differences between Dutch and
Swedish cannot be attributed to the same linguistic and historical factors as for the
High and Low German loans. The contact with French was more intense in the
directly bordering Dutch-speaking area over a much longer period (from the early
Middle Ages on) than in the non-bordering Swedish-speaking area, where French
words mostly entered in the high-prestige times of French as a court language in the
seventeenth and eighteenth centuries. The long-lasting contact situation of Dutch
resulted in a considerably larger number of loanwords (63.8%) compared to Swedish
(14.6%).
The long-term language contact with French brings about some methodological
problems for determining Latin loans in Dutch. In Dutch, many Latin words were not
borrowed directly but via French. Direct borrowing from Latin was more frequent in
Swedish. As our data pertain to the direct loan-giving language, Latin words which
were imported via French were counted as loans from French. For example, a loan
like Dutch civiel / Swedish civil counts as a French loan (of Latin origin) in Dutch,
17
whereas it was categorized as a Latin loan in Swedish. The main cause of the higher
number of Latin loans in Swedish is, therefore, the route along which borrowing took
place. It should also be noted that in many cases it cannot be established whether a
Latin loan was adopted directly from Latin or via French. In our database the Dutch
dictionary indicated ‘French or (Medieval) Latin’ in approximately 5% of the cases.
In these cases we categorized the words as French loans, which means that there may
have been a slight bias towards French loans in our Dutch database. Many French
loans in Dutch can easily be identified to be part of the “Euro-Latin” used in the
formal speech of many European languages, and are identical to the corresponding
Swedish Latinisms.
Considering these facts, we decided to add up the Latin and French loanwords,
thus expressing the number of words which came in via Latin or French. This allows
a valid comparison of the most important routes for the borrowing of Romance words
into both Germanic languages considered in our study. With 86.4%, the number is
46.6% higher in Dutch than in Swedish, with 39.8%. This suggests that overall the
influence of Romance languages has been considerably higher on formal Dutch than
on formal Swedish.
In Figure 5 our results of the origins of loanwords are compared with the findings
reported by Van der Sijs (1996) and Gellerstam (1973). We have taken Low German
and High German together, since the previous investigations do not make this
distinction either. Similarly we have combined French and Latin to form a new
category, which we refer to as Romance.5 The comparison shows a striking
5 Note that Van der Sijs includes Italian and Spanish in this category and Gellerstam includes
Greek.
18
resemblance. Apparently, the percentages of words that Dutch and Swedish borrowed
from different languages are very robust and insensitive to differences between the
corpora.
Figure 5 here
Finally, in Figure 6 we consider the percentages of loanwords from specific loan-
giving languages in relation to the whole lexicon, including inherited words. The
figure shows clearly that the Low German loans are the only reason why the number
of loanwords is higher in Swedish than in Dutch. If we left the words of Low German
out of consideration, the number of loanwords would be about equal for the two
languages (see dotted line in Figure 6). Considering the fact that for methodological
reasons we are unable to assess the exact size of the influence of Low German on
Dutch, and speculating, furthermore, that the influence of Low German on the two
languages may have been rather similar, the contrast between the two languages may
not be as striking as suggested above. What remains, however, is the greater influence
of Romance on Dutch and the greater influence of High German on Swedish.
Figure 6 here
19
6. Conclusions
Our comparison of similarly constructed samples of formal spoken Swedish and
Dutch revealed some clear differences in the composition of the lexicons, in spite of a
common genetic background and parallels in the history of language contact. Swedish
has many more loans than Dutch, namely 44.4% against 27.9%. Moreover, the nature
of the loanwords differs. Within the Swedish loans there is a large compartment of
Low German (38.7%), whereas most loans in Dutch have a French origin (63.8%).
The differences in terms of the number and distribution of loanwords between the
lexical profiles of Swedish and Dutch appear to be stable, as they were attested both
in the present study and in previous studies. They can be attributed to differences in
the linguistic distances between source and borrowing languages and to differences in
the intensity of the contacts.
The higher number of French loans in Dutch can easily be explained. The contact
between Dutch and French was more intensive and widespread than between Swedish
and French. The higher number of Low German loans in Swedish needs some
clarification. The intensity of the contact of Swedish and Dutch with Low German
was almost identical. One would therefore expect similar numbers of Low German
loans in the two languages. This is not what we found. We explained this by the fact
that Low German loans are less easily identifiable in Dutch than in Swedish due to the
larger similarities between the lexicons of Middle Low German and Middle Dutch.
The language contact probably resulted in frequency shifts and semantic adaptation of
inherited words rather than a larger incidence of loans.
20
It should be noted that the reported findings pertain to formal speech. Most
probably the speeches and debates in the European Parliament which formed the basis
of our study were well prepared and can therefore not be characterized as
spontaneous. In this sense, the nature of the Europarl sample may be very close to the
newspaper texts analyzed in previous studies of Dutch and Swedish. As far as we
know, there have been no studies of the lexical profiles of informal spontaneous
speech. To fill this gap we analyzed (part of) the speech produced in the component
‘face-to-face interactions’ of the Corpus Spoken Dutch, collected between 1998 and
2004 (Van Bezooijen, Gooskens & Kürschner, fc.). These are conversations about
everyday topics between friends and relatives, recorded at home without an
interviewer being present. The speakers are from various regions in the Netherlands
and Flanders, of both sexes and of different age groups. To optimize the comparison
with the findings of the present study, we applied the same selection procedure and
coding. The distribution of the loanwords in this spontaneous corpus is very similar to
the distribution found in the Europarl corpus. This confirms that the lexical profiles of
Dutch are very stable and independent of the level of formality, at least as far as the
most frequently used words are concerned.6
6 We thank the editors and an anonymous reviewer for valuable comments on earlier versions of
this paper.
21
References
Allén, S. 1970. Nusvensk frekvensordbok baserad på tidningstext. 1. Graford,
Homografkomponenter. Stockholm: Almqvist & Wiksell.
Bezooijen R. van, Gooskens, C. & Kürschner, S. Forthcoming. Wat weet de
Nederlander van de herkomst van Nederlandse woorden? Nederlandse Taal en
Letterkunde.
Braunmüller, K. 1995. Semikommunikation und semiotische Strategien. Bausteine zu
einem Modell für die Verständigung im Norden zur Zeit der Hanse.
Niederdeutsch und die skandinavischen Sprachen II. K. Braunmüller (ed.), 35-
70. Heidelberg: Winter.
Braunmüller, K. 2004. Niederdeutsch und Hochdeutsch im Kontakt mit den
skandinavischen Sprachen. Eine Übersicht. Deutsch im Kontakt mit
germanischen Sprachen. H. Haider Munske (ed.), 1-30. Tübingen: Niemeyer.
Diercks, W. 1993. Zur Verwendung prä- und postmodifizierender Morpheme im
Mittelniederdeutschen und in den skandinavischen Sprachen. Niederdeutsch
und die skandinavischen Sprachen I. K. Braunmüller & W. Diercks (eds.), 161-
194. Heidelberg: Winter.
Edlund, L.-E. & Hene, B. 2004. Lånord i svenskan: om språkförändringar i tid och
rum. Stockholm: Norstedts.
Gellerstam, M. 1973. Etymologiska frekvenser i det centrala ordförrådet.
Folkmålsstudier 23: 70-78.
22
Goossens, J. 2000. Herauslösung und Herausbildung des Niederländischen.
Ausgewählte Schriften zur niederländischen und deutschen Sprach- und
Literaturwissenschaft. J. Goossens (ed.), 197-211. Münster etc.: Waxmann.
Hellquist, E. 1980 [1922]. Svensk etymologisk ordbok. Tredje upplagan. Malmö:
Liber. Also available at: http://runeberg.org/svetym/.
Sijs, N. van der. 1996. Leenwoordenboek. De invloed van andere talen op het
Nederlands. Den Haag: Sdu / Antwerpen: Standaard.
Sijs, N. van der. 2005. De geschiedenis van het Nederlands in een notendop.
Amsterdam: Bert Bakker.
Thomason, S.G. 2001. Language contact. Edinburgh: Edinburgh University Press.
Veen, P.A.F. van der & Sijs, N. van der. 1997. Etymologisch woordenboek. De
herkomst van onze woorden. Utrecht / Antwerpen: Van Dale Lexicografie.
Wessén, E. 1960. Våra ord. Deras uttal och ursprung. Kortfattad etymologisk ordbok.
Stockholm: Norstedt.
Figure 1. Schematized intensities of language contacts for Dutch (full lines) and
Swedish (dotted lines) between 1000 and 2000; the degree of intensity is indicated by
boldness.
High and Low German
French
Latin
1000 1200 1400 1600 1800 2000
72,1
55,6
27,9
44,4
-
10
20
30
40
50
60
70
80
90
100
Dutch Swedish
inherited words loan words
Figure 2. Percentage of inherited words and loanwords in the Dutch and Swedish
Europarl corpora
27,9
44,4
30,7
47,2
0
10
20
30
40
50
60
70
80
90
100
Dutch Swedish
Europarl previous investigations
Figure 3. Percentages of loanwords in Dutch and Swedish in the present Europarl
corpus and previous investigations (Van der Sijs (1996) for Dutch and Gellerstam
(1973) for Swedish)
Figure 4. Origin of the loanwords (proportions of the total number of loanwords) in
the Europarl corpora
22,6
,3
63,8
9,2
2,1 2,1
25,2
4,6
14,6 14,3
38,7
1,6 1,0
0
10
20
30
40
50
60
70
80
90
100
Latin
Gre
ek
Fre
nch
Hig
h G
erm
an
Low
Germ
an
Englis
h
oth
er
Dutch Swedish
86,4
9,2
39,8
53,0
82,0
7,4
45,3
51,1
-
10,0
20,0
30,0
40,0
50,0
60,0
70,0
80,0
90,0
100,0
Romanic German (L + H) Romance German
Dutch Swedish
Europarl previous investigations
Figure 5. Origin of the loanwords in percentages in the present investigation and in
earlier studies (Van der Sijs (1996) for Dutch and Gellerstam (1973) for Swedish)
Figure 6. The percentages of loanwords from specific loan-giving languages in
relation to the whole lexicon in the Europarl corpora
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Dutch Swedish
inherited
Low German
High German
Romance
Greek
English
other